[common] non-manifold faces

[EricBuijs]

When rendering the common.scad some non-manifold faces are detected but this doesn't appear to be a problem when slicing in Cura. Let's see how printing goes.

[muesli]

@EricBuijs Curious, how did the print go (should you have already finished it)?

I also see the OpenSCAD warning, and that seems to be related to the connectors (it disappears if I disable them).

I'm actually tempted to just give it a go myself, but was still hoping for @Chaos99 to update the connectors for old OpenSCAD and to match the bigger size of the original design - so I could test compatibility with existing old modules as well.

[EricBuijs]

@muesli I'm going to print it later today but I don't think it will be a problem. Here is a link to a screenshot in Cura: https://osada.usezot.net/cloud/eric/common_cura.png

You're correct the non-manifold faces are in the connectors. I checked that with Meshlab.

I'll keep you posted.

[EricBuijs]

@muesli @Chaos99 I finished the first print of the common.scad and it wasn't as easy as I expected. After some trials I decided I needed a brim to get proper adhesion. I think the reason for this that my slicer starts with the connectors which are tiny. I suggest to increase (double?) the size of the connectors since it would certainly help with printing but I think also with attaching the modules smoothly to each other.

link to a photo of the first print

[muesli]

@EricBuijs Nice one! Yeah, as mentioned in the other ticket we'll try to stick to the size of the original connectors. They're fairly easy to print, even though we could reduce the overhang a bit.

[Chaos99]

Possibly fixed this in MR #23
I don't get warnings in my openscad anymore, but please check other versions / renderers / importers.

[EricBuijs]

@muesli: my apologies for asking these stupid questions but how do I download the whole project including the latest pull request from @Chaos99 ?

[Chaos99]

Easy way: browse to my fork on the GitHub page, select the branch that's the source of the pull request, use the download button or checkout with git.
Proper way: add my fork as another remote to your local git repo, then (from your master) do a 'git merge remote/branch --no-commit' (with the proper remote and branch name) to get a preview of the merge.

[EricBuijs]

@Chaos99 @muesli I can confirm that all warnings are gone in OpenSCAD version 2018.03.17 (OSX). Double checked in Meshlab.

[EricBuijs]

@Chaos99 @muesli Printing went very smooth. I used a brim but I don't think that's neccesary. I printed the common.scad two times and tried to fit the connectors but that was hardly possible. My suggestion is to make the male connector a bit smaller both in length and width.

https://osada.usezot.net/cloud/eric/common_Chaos99.jpg

[muesli]

@EricBuijs You can change the width variables in connectors_female and connectors_male. I'm going to try it out with 7 for female and 8 for male now.

[muesli]

@Chaos99 I do still see this particular warning, even in latest OpenSCAD from git master:

"UI-WARNING: Object may not be a valid 2-manifold and may need repair!"

This happens at the end of each render (F6).

[EricBuijs]

@muesli @Chaos99 I can reproduce in OpenSCAD 2015.03 and OpenSCAD 18.03.17 but only with the complete assembly! Base.scad and Common.scad alone don't show any warning when rendering in OpenSCAD 2015.03 and OpenSCAD 18.03.17. Given that the problem is limited to the complete assembly and not the individual parts I think this is not a problem.
EDIT: I initially wrote that this was limited to 2015.03 but I didn't scroll to the bottom of the console in 18.03.17.